If you're combining two sets, and finding the average of the combined set, then you're always calculating what is known as a "weighted average". The average is "weighted" by the sizes of the sets (by the ratio of their sizes, technically). So if one set is twice as big as the other, the average you calculate will be "twice as close" to the bigger set's average -- the bigger set is twice as important as the smaller one.
So if you wanted to find the average income for people in the USA and in Canada, then since the USA has about ten times as many people as Canada, the USA's average is going to be far more important than Canada's average. You could not just find the USA average, and the Canada average, and then average those two numbers. You'd need to do a weighted average calculation. Weighted averages are too big a topic to get into in a short post, but one way to get the right answer is to give the USA average a weight of 10 (i.e. multiply it by 10) and the Canada average a weight of 1, and then divide by 10 + 1 = 11. So if average income in the US is $31,000 and in Canada is $20,000, then the overall average will be
[(10)(31,000) + (1)(20,000)] / 11 = 30,000
If you look at how far the answer here is from the individual averages, you'll notice it's 10 times as far from Canada's average than from the USA average. In non-grammatical speak, you could say the answer is "10 times as close" to the USA average (that's not correct English though). That's identical to the ratio of the sizes of the two populations, and weighted averages always work out that way. That observation is the basis for a different weighted average method that is sometimes called "alligation", which you could look into, though that method takes a bit of practice. If you can learn it though, it's the best weighted average method for GMAT purposes.